<!-- Some styling for better description lists --><style type='text/css'>dt { font-weight: bold;float: left;display:inline;margin-right: 1em} dd { display:block; margin-left: 2em}</style>
qwebirc96010: is it possible for a card to be successfully initialized in '/dev/dvb' with no errors on dmesg/syslog, and yet no capable of scanning any channels due to something faulty inside the ard?
<br> *card
<br> not*
x_martinb: <u>merbanan</u>: hello :) I've seen that you know a lot of details about mn88472 (and you've documented that knowledge here - thanks for that! https://linuxtv.org/wiki/index.php/Panasonic_MN88472)
<br> I wonder if you know the differences to mn88473
<br> (it seems that there are only a few differences, specifically I'm interested in statistics support for the mn88473)
crope: big differences
<br> some registers are same, but most are off by N
<br> <u>x_martinb</u>: if you can, just pressure I can get the documetation
<br> some vendors who are using that chipset are the one who should first pressure
x_martinb: what do you mean exactly with "pressure"? send someone an open letter?
crope: <u>x_martinb</u>: and it is merbanan who makes documentation :)
x_martinb: ah so we ask him nicely if he can add the docs for mn88473, just like he did for ..2? :)
crope: <u>x_martinb</u>: yes, just ask open linux drivers. problem is that developers has many times zero documetation unless device vendor helps
<br> very often the key for documentation / sdk is device vendor who uses those chips...
<br> I ask pretty often docs, but almost never got without any help from device vendors...
<br> <u>x_martinb</u>: actaully, last week I sent new request for the doc (I earlier tried with the help of eu division but failed)
x_martinb: <u>crope</u>: sorry, I have some connection issues today
<br> just to clarify this: does merbanan have access to the docs / vendor sdk (so he can simply write the documentation) or do we have to ask someone else (who would that be) for docs / sdk?
merbzt: <u>x_martinb</u>: I tried to get dox a few years ago
<br> they even called me, but as I was not going to buy anything I didn't get any info from them
crope: <u>merbzt</u>: last week I sent request to that new entitny who owns the chips <name it>
merbzt: <u>crope</u>: yeah, I saw that to, lets hope they are better at releasing info
<br> or I saw that someone else bought them
crope: we talked with EU division "long" about those chips when they were Panasonic
x_martinb: <u>merbzt</u>: so it seems you found out all this by RE'ing the windows driver or analyzing usb captures?
merbzt: <u>x_martinb</u>: there is source code also
x_martinb: interesting, I did not search for mn88472 yet but I could not find anything for the mn88473
merbzt: <u>crope</u>: so you actually got hold of a person ? I only found a sales guy, he was not so helpful
x_martinb: (the latter one is the one I'm interested in, since I have no device with mn88472)
crope: look mn88473 driver - it is more simplifier than mn88472
merbzt: <u>x_martinb</u>: the mn88473 has some register offset, but probably works the same way as mn88472
crope: <u>merbzt</u>: yeh, got the persons who relly understood the issue.... but it was japan hq which denies
x_martinb: <u>merbzt</u>: ok, thanks
merbzt: ahh impressive that you even got so far
crope: <u>merbzt</u>: : ]
merbzt: <u>x_martinb</u>: what do you want to do ? what features are missing
x_martinb: <u>merbzt</u>: I was just interested in some statistics (signal strength, etc.)
<br> so nothing critical
crope: <u>x_martinb</u>: I can likely measure, but it is too much effort to rev engineer so I dont....
merbzt: <u>x_martinb</u>: http://pastebin.com/vhEbKmAb
crope: u know, dumping the resiters against the signal generator....
x_martinb: <u>crope</u>: well, I was hoping to cheat this a bit: using DVB-T and packet capture on windows, then play around with the antenna so signal gets worse -> find out the registers and see if they are similar to ...2
merbzt: <u>x_martinb</u>: if you do the work I can guide you
<br> <u>x_martinb</u>: dvb-t should be doable, we can start with that
<br> dvb-t uses one reg per statistics
<br> dvb-t2 uses load and result mechanism
crope: <u>merbzt</u>: x_martinb signal stengt is just AGC gains. you can do it
merbzt: AGC gain should be easy
<br> https://linuxtv.org/wiki/index.php?title=Panasonic_MN88472#Signal_strength
x_martinb: <u>merbzt</u>: seems like the statistics are sorted by "implementation complexity" :)
crope: <u>merbzt</u>: thats most likely agc feedback :)
merbzt: just dump all registers with cable connected and unconnected
crope: <u>x_martinb</u>: yeh, trivial when you know reisters and scale
x_martinb: <u>merbzt</u>: let's say I start with signal strength: seems there is only one reg for DVB-C/T/T2. do you think registers are the same on ...3?
<br> ah :)
crope: 3*255 register is pretty much o_O
x_martinb: that's an even easier test as the one I came up with :D
merbzt: it should be around -0x1D from 0x8E
crope: <u>merbzt</u>: you dont have the signal generator?
<br> <u>merbzt</u>: good work still!
x_martinb: <u>merbzt</u>: ok, I'll give this a go and let you know if I need further help
<br> thanks for now!
merbzt: I have one but I never had the energy to get everything working
crope: <u>merbzt</u>: as I am too lazy to measure. I have one PCI generator I propably can give to you
merbzt: I also have a computer with a MN88472 MN88473 demod in
<br> if you sort out the agc feedback I'll arrange so you can test on it
<br> <u>crope</u>: like you I have far to much other stuff to do, but what PCI hardware is it you have ?
crope: dektec pci-xxx
<br> without any licenses, just dvb-t and dvb-c
<br> I will give it to someone who really helps linuxtc
<br> just bought new signal genereator 3Gz, so I have now 1GHz gen for sell
merbzt: <u>x_martinb</u>: anyway if you need any help just ping me or email, I'm sure you can get this working with some tinkering
<br> <u>crope</u>: gnuradio is quite capable now also, they have lots of modulators
crope: <u>merbzt</u>: y, but thats hw signal generator HP8657A
<br> I have now 2 generators, SME03 and
merbzt: I have the budget version
crope: I have now 2 generators, SME03 and HP8657A
<br> SME03 with all those GSM and tetra options none wants, old nokia
merbzt: http://www.ebay.com/itm/172152204149?_trksid=p2060353.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT
crope: good option if better not avalaible
merbzt: I used it to test the r820t tuner, > 1500MHz you get nothing even if the tuner can lock on the test hardware
crope: but how to genrate acutal signal?
<br> aah, sorry stubid q
<br> <u>merbzt</u>: is it working? lookin one hw with my prof equip it is off to 100kHz ion 1500
merbzt: I wanted to see if there was a slightest chance in hell if I could get the rtl-sdr to receive 1900MHz
<br> I got the tuner to lock at 1893MHz
<br> but the spectrum was dead silence
<br> so I checked how low I had to go and it was around 1500-1600MHz before I picked up the adf4350 generated signal
crope: not sure what my stick locks, but 1500e there is currently almost 0 sinal (not sure how much I have to put tx in able to see it)
merbzt: I had the cable connected directly through
crope: fact the chip is hard to configure....
merbzt: prog has written a "better" driver for the airspy
crope: havent looked that
merbzt: but I doubt there is much to be done for the tv use case of the Astrometa sticks
crope: but in any case kernel driver should support dvb-t and dvb-c to it is not easy
merbzt: as long as you have a good signal it seems to work fine now
<br> the only issue I have heard is when people use strange frequencies
<br> for dvb-c
<br> and dvb-c/annex c is not tested/supported
<br> same with 5MHz/1.7MHz dvb-t2
crope: I think you break the high freq dvb-c :DDDD
merbzt: yeah maybe need to switch the antenna
<br> from cable to air
crope: fortunately it is not regression as that patch went in that early (before demod)
merbzt: but it is hard without a proper way to test it
crope: yep
<br> and I am too lazy... :)
<br> I have too many things to test, so I am not able to test all
merbzt: yeah, you are too lazy, only wrote like half of the drivers in recent years
crope: tuner are evil
<br> +s
merbzt: I still have the dibcom and connexant mn88472 dongles to write support for
crope: most worst are DVB-S, I dont have any real signals
merbzt: modulate your own
<br> I think an rpi can do that now
crope: earlier I has remote computer on pctv on .de but not anymore :S
merbzt: I have access to 10 satellites
crope: I have only generator....
merbzt: I can give you access if you need
crope: what kind of hw? Can I switch USB on the fly?
<br> sure I would like too
merbzt: well what kind of hardware do you need ?
crope: pctv there was many satellites and usb power bar I could change devices
<br> currently not any...
merbzt: so you want to be able to enable/disable usb ports ? with connected hardware to
crope: because those devices were connected to usb ports
<br> so I was able to reboot computer, reconnect USB
<br> select satellite using diseq
merbzt: need to find that kind of usb hub then
crope: hub that could be remotely controllrd
<br> but really, I am not going to work with DVB-S anymore
<br> generatot is enough
merbzt: yeah, it is so much easier to work with everything locally
<br> software controllable hub $300.00
<br> a little bit to much